JP2008015722A - Data processing system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To efficiently monitor faults of a sensor node with high reliability on the assumption of intermittent operation while saving and effectively utilizing limited resources of the sensor node. <P>SOLUTION: An execution timing of fault determination processing of the sensor node is determined in accordance with a preliminarily set rule on the basis of an intermittent operation interval supposed in the sensor node. That is, when a change of the intermittent operation interval of the sensor node is requested, a monitoring part of a system changes the execution timing of fault determination processing so as to make the execution timing follow the change of the operation interval, immediately after confirming the operation completion of the change of the operation interval of the sensor node. In the case that intermittent operation intervals required per sensor node or classification of sensor nodes are different, execution timings of fault determination processing are properly used per sensor node or classification. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a technique for monitoring the operating states of a large number of terminals connected to a network.

  In recent years, a network system (hereinafter referred to as a sensor network system) in which a small electronic circuit having a wireless communication function is added to a sensor and various information in the real world is taken into an information processing device in real time has been studied. A wide range of applications are considered for sensor network systems. For example, temperature and humidity, acceleration, etc. are measured by a small information terminal (hereinafter referred to as a sensor node) that integrates wireless circuits, processors, sensors, and batteries. The monitoring results are sent to the server, etc. by wireless communication, and the activity status of people and the storage of things based on the monitoring results collected by the server. A technology that analyzes the state and links it to a service has been proposed.

  A sensor node equipped with a power source such as a wireless communication circuit, a processor, a sensor, and a battery has characteristics that are not found in a conventional computer system. The sensor node's communication and power supply are wire-free, which makes it possible to observe the state of things and places that could not be observed with the sensor network system. In order to put this sensor network system into practical use, there is a sensor node that can continue to transmit sensing data for a long time without limited energy constraints (for example, with fewer battery replacements). Desired. In addition, in order to continue to provide some service using sensing data, efficient operation management of the entire sensor network system including a plurality of sensor nodes is required.

  In particular, whether each component of the system is in an abnormal state (whether a failure has occurred or no indication of failure), and vice versa Monitoring is required as much as possible. For example, there is a technique in which each client detects the presence or absence of a failure based on a hard beat signal transmitted from a server as a failure detection means of a system component (see, for example, Patent Document 1).

JP-A-7-56838

  It is desired that the sensor net sensor node targeted by the present application operate with as low power consumption as possible. Therefore, the sensor node often suppresses the power consumption significantly by performing an intermittent operation that repeatedly starts and pauses at certain time intervals. This intermittent operation may vary depending on a request from an application that uses data. For example, different data update frequencies may be requested according to the type of sensing data, or different data update frequencies may be requested according to the sensing time zone even for the same sensor. More specifically, when considering an application for air conditioning control by monitoring indoor temperature / humidity, the frequency of temperature / humidity measurement and data transmission is reduced during periods when there are no people in the room (for example, at night). (For example, once every 5 minutes), it is conceivable to increase the temperature and humidity measurement and the data transmission frequency (for example, once a minute) during the time zone (daytime) when a person is present in the room.

  Furthermore, from the point of view of system operation management, it is necessary to adaptively adjust the failure determination execution timing of the sensor node as the data transmission interval changes. If the execution timing of the failure determination processing of the sensor node is not appropriate (for example, fixed), inefficient determination processing will be executed after the change of the operation interval, or the failure detection delay will exceed the allowable range. Problems arise. On the other hand, it is unrealistic from the viewpoint of computer resource consumption of the system to always execute the failure determination process repeatedly.

Also, in a system that does not assume that the transmission interval is changed at all, there may be a problem that erroneous detection or failure of detection may occur in failure monitoring.
Furthermore, it is very inefficient to deal with the above problem by individually setting the execution timing of the failure determination processing by the system administrator, particularly when the number of sensor nodes is increased.

  Therefore, the present invention has been made in view of the above problems, and efficiently and efficiently monitors the failure of a sensor node on the premise of intermittent operation while saving and effectively using limited resources of the sensor node. The goal is to achieve high reliability.

In order to solve the above problems, the outline of the representative invention disclosed in the present application is as follows.
A communication unit that receives data transmitted at a predetermined interval from a terminal connected via a network, and a determination unit that performs abnormality determination of the terminal with or without data to be received, A data processing system for determining an interval for performing the abnormality determination for the terminal based on the predetermined interval according to a predetermined rule.

Therefore, according to the present invention, the execution timing of the failure determination processing of the sensor node can be optimized, and actual failure detection can be effectively realized. In addition, since the execution timing of the failure determination process is changed according to a rule that can be commonly applied to all or a plurality of sensor nodes included in the system, there is no need for setting. Therefore, it contributes to the realization of efficient system operation and maintenance, and the quality of services provided using the system can be kept high.
For example, there is an effect of suppressing the consumption of energy (power supply) by suppressing the operation frequency of the sensor node in a time zone when frequent air conditioning control is not required. In addition, depending on the radio frequency band used, it may be possible to improve the reliability of radio communication by increasing the transmission frequency when the person becomes an obstacle during the day and the quality of the radio communication deteriorates. Furthermore, by making this possible from a remote location, it is possible to improve the quality of system operation services at low cost.

Hereinafter, an embodiment of the present invention will be described with reference to the accompanying drawings.
FIG. 1 is a block diagram illustrating an example of a sensor network system (hereinafter referred to as a sensor network system) to which the present invention is applied according to the first embodiment. In the illustrated sensor network system, sensing data from the sensor nodes SN1 to SNn are transferred to any one of the base stations GW1 to GWm by wireless communication, and the base station GW transmits the sensing data on the wireless communication to the wired network WDN. Act as a gateway to The sensing data transferred by the base station GW is sent to the sensor network management server SNMS via the wired network WDN and used by the user terminal UST.

The sensor nodes SN1 to SNn that are one of the components of the sensor network system will be described.
In FIG. 1, sensor nodes SN1 to SNn are sensor nodes that output sensing data or a preset ID (identifier) by wireless communication. These sensor nodes SN1 to SNn perform wireless communication with the base station GW through the wireless network WLN. The sensor nodes SN1 to SNn are attached to a predetermined part of a person, for example, for the purpose of monitoring the state of the person. Each sensor node SN1 to SNn transmits sensed temperature, pulse, acceleration, and other data to the sensor network management server SNMS via the base station GW.

  SN1 to SNn are sensor circuits SDEV for measuring (sensing) a wireless circuit RFC that realizes wireless communication, a person who is in contact with or around, a physical object or something (for example, temperature, acceleration, pressure, etc.) Microcomputer CPU for executing an actuator device ADEV that causes some physical phenomenon (for example, light emission, character display, vibration, etc.) to act on a person, thing, or environment that is in contact with the surroundings, or a sensor node itself. , Comprising a power supply PWR for driving the above components. For example, when the power supply PWR is a battery, the power that can be used is limited. In order to extend the life as much as possible, it is necessary to control not to supply power to the components that are not used during the time when the sensing process is not necessary, while suppressing the power consumption during operation such as the sensing process.

  The microcomputer CPU of the sensor nodes SN1 to SNn performs the operation of the sensor control unit SCTL that manages and controls the operation of the sensor device SDEV provided in each sensor node SN1 to SNn, and the operation of the actuator device ADEV provided to the sensor nodes SN1 to SNn. Actuator control unit ACTL that manages and controls, wireless communication control unit RFCTL that manages and controls wireless communication with other sensor nodes SNi or base station GW via wireless network WLN, and power supply PWR provided in sensor nodes SN1 to SNn And a power control unit PWCTL that controls power supply from the power supply, and a sensor control unit, an actuator control unit, a wireless communication control unit, and an operation control unit MCTL that controls the power control unit. The sensor control unit, actuator control unit, wireless communication control unit, power supply control unit, and operation control unit can be realized as programs on the microcomputer CPU.

  The sensor devices included in the sensor nodes SN1 to SNn are, for example, a temperature sensor, a humidity sensor, an acceleration sensor, a pressure sensor, and a pulse sensor, and have an identifier for identifying an individual or an individual object. .

  The actuator devices included in the sensor nodes SN1 to SNn include LEDs that are light emitting elements, a liquid crystal display device LCD, and the like. For example, a sensor node in the form of a wristwatch that includes them can be realized. Note that the actuator control unit ACTL is not provided in the sensor nodes SN1 to SNn having no actuator.

The base station GW that is one of the components of the sensor network system will be described.
The base station GW manages IDs, a wireless communication control unit GRFC that controls wireless communication with a plurality of subordinate sensor nodes SN1 to SNn, a wired communication control unit GNIC that controls wired communication with the sensor network management server SNMS, and IDs ID management unit GIDM, command management unit GCMM for managing commands, upstream packet conversion unit UPC for converting packets transmitted and received between the sensor network server SNMS and sensor nodes SN1 to SNn, and similarly converting downstream packets A downstream packet converter DPC. The base station GW manages one PAN (Personal Area Network) and can transmit and receive data by wireless communication with an arbitrary number of sensor nodes belonging to the PAN.

  The ID management unit GIDM assigns a local ID to a sensor node that newly requests to participate in the PAN, converts a local ID to a global ID, or converts a global ID to a local ID. It has an ID management table for this purpose. The local ID is an ID having a scope within one PAN (Personal Area Network). Usually, one base station GW having a function of managing a local ID exists in one PAN. The local ID has a shorter bit length than the global ID, and can be expected to reduce the power consumption of wireless communication when included in a transmission / reception packet.

  The global ID is an ID that allows at least an application on the sensor network system or an upper system of the sensor network system to identify the sensor node. Since one sensor network system may include a plurality of PANs, a sensor node is managed using a global ID in an application or a host system of the sensor network system. For this reason, the global ID has a larger number of bits than the local ID. For example, in the ucode of the ubiquitous ID center, the basic consists of 128 bits, and the EPC global EPC (Electronic Product Code) consists of 96 bits. On the other hand, the local ID is about 16 bits, for example.

  The downstream packet is a command for the sensor node. A command for changing an operation parameter such as an operation interval of a sensor node, a command for inquiring a current set value of a certain operation parameter, and the like. On the other hand, the uplink packet is observation data by the sensor node or a command response to the sensor node.

  The command management unit GCMM transmits the data transmitted from the wireless communication control unit GRFC to determine whether the upstream data from the sensor nodes SN1 to SNn to the sensor network server SNS is a response to the command transmitted to the sensor node. Manage the commands When a command is transmitted from the downstream packet conversion unit DPC toward the sensor nodes SN1 to SNn to the subordinate sensor nodes SN1 to SNn, the command is registered.

A wireless network WLN that is one of the components of the sensor network system will be described.
The wireless network WLN is constructed by a wireless method with low power consumption. For example, a standard such as IEEE.802.15 is used. However, the present invention is not limited to this standard.

A user terminal that is one of the components of the sensor network system will be described.
The user terminal UST is an application that applies a sensor network management client SNMC or a sensor network system for a system administrator of the sensor network system to perform operations such as system state monitoring and operation parameter confirmation / setting. It consists of an application client APPC that monitors the state of the environment and provides a specific service based on the monitored result. However, the sensor network client SNMC and the application client APPC may be physically on the same computer as the sensor network management server or on a different computer. Each may be on a different computer.

The sensor network management server SNMS, which is one of the components of the sensor network system, will be described.
The sensor network management server SNMS manages sensing data collected by a plurality of base stations GW to GWm via a wired network WDN (for example, the Internet), and sense data in which semantic information is added to a user terminal (user computer) UST. I will provide a.

  The sensor network management server SNMS has the following processing units. A wired communication control unit SNIC that communicates with the base station GW, the user terminal UST, and an external device via the wired network WDN. A session management unit SSM that controls and manages transmission / reception with the user terminal UST. An event processing engine EVE that receives received data as an event and activates a predetermined action based on a result of a condition determination according to a pre-installed logic. A script management unit SCM that receives and manages logic incorporated in the event processing engine from a user terminal in advance as a script. An external device cooperation adapter EXTIF that operates an external device as one of the actions that the event processing engine EVE starts. A timer TMR that issues a timer event that is one of the events received by the event processing engine at a timing specified in advance. Sensor network configuration for integrated management of configuration information (ID, installation location, operating status, etc.) of sensor nodes SN1 to SNn and base stations GW1 to GWm and sensing data mounted on each sensor node to be managed Management department CFM. A sensor network command management unit CMM that issues and manages commands to the base station GW and the sensor nodes SN1 to SNn from the user terminal UST or the like, or from the event execution engine or the sensor network configuration management unit. A database management unit DBMS that manages data handled by the sensor network configuration management unit and the session management unit as a database.

  If the data from the base stations GW1 to GWm received by the wired communication control unit SNIC is a response to a command issued by the sensor network command management unit CMM in advance, the event processing engine EVE transfers it to the sensor network command management unit CMM However, if it is sensing (observation) data, the corresponding action is activated according to the logic incorporated in advance. For example, if the received data is sensing data in which an abnormality monitoring logic has been incorporated in advance, it is determined whether or not the value is abnormal. In addition to updating, if a predetermined action is registered in advance, the action is activated. The predetermined action is, for example, an operation in which a message indicating that an abnormality has been detected is displayed on the application client APPC on the user terminal UST. In response to an instruction from the user terminal UST, the script management unit SCM is set in advance. Set through. In addition, the event processing engine EVE may activate a specific operation triggered by a timer event from the timer TMR. For example, every time a periodic timer event is received from the timer TMR, the event processing engine EVE determines whether or not a failure has occurred, including a life / death determination of a sensor node that is a component of the system. Execute the process.

  The sensor network configuration management unit CFM manages operation parameters, operating states, and the like of each device configuring the sensor network system such as the base station GWi and the sensor node SNi. In addition, an interface related to registration and search of each device is provided to the user terminal UST and the like. In FIG. 9, items to be specifically managed (FIG. 9A, base station management items, (B) sensor node management items), base station management table (FIG. 9C), and sensor node management table (FIG. 9D )) An example is shown. Some management items (or attributes) are items that are not basically changed during operation, such as “name” and “MAC address”, and may be changed during operation, such as “data transmission interval”. There are certain operation parameters or items that are automatically updated according to the situation during operation, such as “connection node list”, “operation status”, “temperature latest value”, and “last data reception time”. The update process for automatically updated items is realized, for example, when the event processing engine EVE calls the attribute network update process of the sensor network configuration management CFM. Further, as in the “connection node list” and “belonging base station ID”, the connection relationship such as which base station GWi belongs to each sensor node SNi is managed as configuration information. For example, when it is desired to send a command (command) for changing an operation parameter (sensing interval, data transmission interval, etc.) to a certain sensor node SN based on this connection relationship, the command is first transmitted to which base station GW. The transmission path is determined whether or not (request for command transfer to the sensor node) is to be performed.

Note that one of the base stations GW may physically have a sensor network management server function.
The wired network WDN in FIG. 1 is not necessarily a wired network and may be a wireless network. However, the wireless network WLN forming the wireless sensor network in FIG. 1 does not need to be a special wireless network that requires low power consumption. For example, a so-called wireless LAN such as IEEE.801.11a / b / g is used. good.

An operation state monitoring of the sensor node in the sensor network system will be described.
Now, considering the continuous provision of some service using the system, system operation management is essential. One of the operations management is the task of monitoring various system abnormalities. Even in the sensor network system, for example, it is necessary to comprehensively grasp whether or not a large number of sensor nodes are operating normally, and if there is an abnormality, it is necessary to detect it quickly and take a possible countermeasure.

  As one of the means to monitor the abnormality of a device that is operating remotely, the management server that is the monitoring device sends a command to inquire about the status of the device to be monitored at an arbitrary timing, and the response is relatively instantaneous Whether or not the device to be monitored is operating normally can be grasped by checking whether or not it returns during a specific waiting time. However, in the case of a sensor network system, the sensor node is intermittently operated so as to repeat operation and pause at certain intervals to reduce power consumption, so it is difficult to respond instantaneously at any timing. .

As another means, a device to be monitored sends a specific heartbeat signal (a signal for determining whether it is alive or not) at a fixed interval, and the management server that is a monitoring device determines whether or not the heartbeat signal is received. There is a method of determining the operating state of the device to be monitored based on the above. That is, the time when the heartbeat signal should be received next is obtained from the time when the heartbeat signal was last received and the heartbeat signal transmission interval, and it is determined whether the time is before or after the current time. If it is, it is judged as normal, and if it is before (in the past), it can be judged as a failure because the last heartbeat signal has not arrived. Even in the case of sensor nodes, the operating state can be monitored by this method. Here, it is desirable to avoid sending a specific heartbeat signal only for monitoring in a sensor node that needs to reduce power consumption. When it is required to periodically sense (observe) environmental information (for example, temperature) using a sensor, the sensing data is periodically transmitted. It can be regarded as a beat signal.
In order to increase the reliability of sensing data acquisition even at the expense of power consumption, there may be an exceptional case where a specific (dedicated) heartbeat signal is frequently transmitted at intervals shorter than the sensing data. . In addition, a sensor node that is supposed to send data irregularly triggered by a mechanical switch, such as a person or object placed or touched by a person or object, also makes a separate life / death determination. For this purpose, heartbeat signals may be transmitted at certain intervals. In this case, the signal to be transmitted may be a signal dedicated to the heartbeat or may include some sensing data.

  2 and 3 are sequence diagrams for explaining the monitoring of a failure in which measurement data of a sensor node does not reach.

  FIG. 2 is a sequence diagram showing an embodiment of sensor node failure monitoring, in which a data reception asynchronous timer (detailed later) is used. The sensor node SNx in FIG. 2 corresponds to one of the sensor nodes SN1 to SNn in FIG. 1, and the timer TMR is the same as that in FIG. 2 corresponds to the sensor network configuration management CFM and the event processing engine EVE in FIG. 1, and the management client SNMC in FIG. 2 corresponds to the sensor network management client SNMC in FIG. In FIG. 2, the base station GW in FIG. 1 is omitted.

  As a premise, in the sequence example of FIG. 2, when the event processing unit EVE receives a timer event, the event processing unit EVE executes a failure determination of the sensor node by using it as a trigger. Here, the sensor node which is a target for determining the failure is “operation state” of “operation” (or “subscription”) at that time in the sensor node management table (FIG. 9D) in the configuration management unit CFM. Are managed as sensor nodes. Basically, the event processing unit EVE executes failure determination processing for every sensor node that is a failure determination target each time a timer event is received.

  In addition, the management client SNMC can be registered in advance so that the configuration management unit CFM is notified when any attribute managed by the configuration management unit is updated. In addition, the management client SNMC sends an arbitrary management attribute (for example, the operating state of the sensor node) to the configuration management unit at an arbitrary timing regardless of whether or not the update notification is registered in advance in the configuration management unit CFM. You can also inquire. The above premise applies not only to FIG. 2 but also to FIGS.

  In FIG. 2, the sensor node SNx joins or associates with the PAN, operates intermittently at regular intervals, senses an observation target during operation, and transmits the observation data by wireless communication. Repeat at regular intervals. When the event processing engine EVE receives the event notification that the sensor node SNx has joined, the configuration management unit CFM sets one of the attributes related to the sensor node SNx, which is a part of the configuration management information (FIG. 9), to “subscription”. To the value "". This update to “subscription” is notified as shown in FIG. 2 if the management client SNMC is registered so as to notify that in advance. FIG. 2 shows a sequence in which the timer TMR is activated by receiving a subscription event from the sensor node, but this process is not necessary if the timer TMR has already been activated.

The basic operation after the sensor node is activated will be further described with reference to FIG.
Whenever the configuration management unit CFM receives measurement data at regular intervals from the sensor node SNx via the event processing unit EVE, the “operation state” which is one of the attributes representing the configuration information is “operation” representing the normal state. If it is not “operating” (for example, “subscription”, “failure”, “unknown”), it is updated to “operating”. Furthermore, another attribute “last data reception time” is updated.

  Every time the event processing unit EVE receives a timer event from the timer TMR, the event processing unit EVE executes an operation state determination (failure determination) process, and all the sensor nodes including the sensor node SNx are processed by the configuration management unit CFM in FIG. In this way, it is determined whether or not a time longer than the “data transmission interval” of each sensor node has elapsed since the “last data reception time” of each sensor node managed in association with each node ID. If there is a sensor node that is operating, it is determined as “failure”. If it is determined as “failure”, the value of the attribute “operating state” of the corresponding sensor node is updated to “failure” in the table of FIG. 9D of the configuration management unit CFM. In FIG. 2, the operating state determination result when the first to fourth timer events are received is normal, and the “operating state” remains “operating”. When the fifth timer event is received, the third measurement data has not arrived from the sensor node SNx for some reason, and is determined as “failure”. This determination result is also notified to the management client SNMC if notification is set in advance. When a failure is detected, the “operation status” is updated to “failure”. However, if the recovery is confirmed after receiving the measurement data, the “operation status” is returned to “operation” again. If the notification setting is registered, “recovery” is notified to the management client SNMC. The values that the attribute “operating state” such as “subscription”, “operation”, and “failure” take are not limited to these, and the expression may be changed, or a variety of possible values may be added.

  When the data reception asynchronous timer shown in FIG. 2 is used, in order to determine a failure at the timing when the event processing engine EVE receives a timer event issued from the timer TMR at a preset timing, FIG. As shown in the figure, detection delay and notification delay are unavoidable, and it is necessary to determine and set the timer event issuance interval after examining the allowable delay time range in advance.

  FIG. 3 is a sequence diagram showing another example of sensor node failure monitoring, in which a data reception synchronous timer (detailed later) is used. In this sequence example, an operation cycle is set in the timer so as to synchronize the transmission timing of the transmission data from the sensor node and the failure determination execution timing as much as possible. The relationship between each part in FIG. 3 and FIG. 1 is the same as that in FIG.

  In FIG. 3, the sensor node SNx joins or associates with the PAN, operates intermittently at regular intervals, senses an observation target during operation, and transmits the observation data by wireless communication. Repeat at regular intervals. When the event processing unit EVE receives the notification that the sensor node SNx has joined, the event processing unit EVE sets the timer TMR to a time interval based on the measurement data transmission interval assumed by the sensor node SNx (for example, an assumed “data transmission interval”). Set to issue a timer event. The configuration management unit CFM updates the “operation state” of the corresponding sensor node SNx to “join” in the sensor node management table (FIG. 9D). This “subscription” is notified as shown in FIG. 3 if the management client SNMC is registered so as to notify that in advance.

  Subsequently, when the configuration management unit CFM receives measurement data from the sensor node SNx at regular intervals via the event processing unit EVE, the configuration management unit CFM holds the attributes (or management items) held in the sensor node management table (FIG. 9D). ) Is “operating” indicating a normal state, and if it is not “operating”, it is updated to “operating” and at the same time in the sensor node management table (FIG. 9D). Update the “last data reception time” stored in. At the same time, the event processing engine EVE issues a restart instruction to the timer TMR, temporarily cancels the timer event issue, and resets the timer event issue under the same conditions. Here, the timer TMR is reset, but this is a problem of the accuracy of the timer event. In order to maintain accuracy so as to synchronize with measurement data, it is desirable to synchronize every time measurement data is received. However, if the synchronization timing can be maintained without resetting by some means, the timer TMR is reset. Is not required.

  In FIG. 3, a scenario is assumed in which the third measurement data has not arrived from the sensor node SNx for some reason. Therefore, a timer event is issued from the timer immediately after the third measurement data is scheduled to arrive. The event processing unit EVE can detect the failure by determining the state of the sensor node SNx as “failure” only by receiving this timer event. More precisely, from the “last data reception time” regarding a certain sensor node SNx held in the sensor node management table (FIG. 9D) in the configuration management unit CFM, the same is also held in the sensor node management table. It may be evaluated whether or not a time longer than the “data transmission interval” has elapsed, and if it has elapsed, it may be determined as “failure”. Here, as a determination result of failure detection, the value of the “operation state” of the corresponding node in the sensor node management table (FIG. 9D) is updated to “failure”. In any case, there is almost no delay in detecting the failure, which is a feature when the data reception synchronous timer is used.

  Further, this determination result can be notified to the management client SNMC, but there is almost no delay in the notification.

  Next, the timer TMR will be described. As shown in FIGS. 2 and 3, there are two ways to use the timer for determining the execution timing of the failure determination process. However, since each has advantages and disadvantages, it is desirable to use them according to the situation.

  The “data reception asynchronous timer” used in FIG. 2 is a timer that is used to continue issuing timer events at an appropriate interval regardless of the data reception timing. One timer is used to start failure determination of multiple sensor nodes. Share with. According to this method, it is necessary to perform determination processing for all monitored sensor nodes sharing a timer at the time of one failure determination execution, and the failure detection delay is determined according to the sensor node having the shortest transmission interval. The amount of calculation tends to increase, and there is clearly a delay time for fault detection (and fault notification). On the other hand, since the timer is shared, the computer resources (mainly memory usage) for implementing the timer are small. Therefore, it is effective when the allowable delay time is relatively loose and when the number of sensor nodes to be monitored is large.

  The “data reception synchronous timer” used in FIG. 3 is a timer that is activated in response to data reception from the sensor node, and is used to issue a timer event after an elapse of an assumed transmission interval + α (for example, 1 second). A timer is prepared for each sensor node, and a timer event according to the sensing data transmission timing of each sensor node is used.

  According to this method, the computer resources (mainly memory) for mounting the timers are increased by the number of sensor nodes. On the other hand, since the execution of failure determination is performed on a sensor node basis, the required minimum amount of computation (time) is sufficient, and the delay time for detecting failure (heartbeat failure) is reduced. Therefore, it is effective when the allowable delay time is severe and the number of sensor nodes to be monitored is small.

Next, the change of the data transmission interval from the sensor node will be described.
The transmission interval of sensing (measurement) data regarded as a heartbeat signal is not necessarily fixed. The transmission interval of sensing data transmitted by the sensor node in accordance with the intermittent operation may vary depending on a request from an application that uses the data. For example, different data update frequencies may be requested according to the type of sensing data, or different data update frequencies may be requested according to the sensing time zone even for the same sensor.

  Note that the transmission interval may be changed by an instruction from the application or sensor network system administrator, or when the data that the sensor node itself observes meets certain conditions, The data transmission interval may be changed autonomously, or the measurement interval or data transmission interval of another sensor node when the measurement result of a certain observation target by a certain sensor node satisfies a certain condition It may be changed automatically in the network management server SNMS).

  Here, from the viewpoint of system operation management, as the data transmission interval is changed in a certain sensor node, it is necessary to adaptively adjust the execution timing of failure determination of the sensor node. If the execution timing of the failure determination processing of the sensor node is not appropriate (for example, fixed), inefficient determination processing will be executed after the change of the operation interval, or the failure detection delay will exceed the allowable range. Problems arise. On the other hand, it is unrealistic from the viewpoint of computer resource consumption of the system to always execute the failure determination process repeatedly. Also, in a system that does not assume that the transmission interval is changed at all, there may be a problem that erroneous detection or failure of detection may occur in failure monitoring.

(Explanation of rules)
In order to solve the above problem, in the present invention, the execution timing of the failure determination process of the sensor node is determined based on the intermittent operation interval (or data transmission interval) assumed in the sensor node according to a preset rule. To do. When using a data reception synchronous timer, for example, the execution timing of a failure determination process of a certain sensor node, and for the first timing, the “data transmission” set in the sensor node after receiving the previous measurement data A rule is set in advance so that the interval is set to a time obtained by adding a margin interval of 1 second, and then executed at the same time interval as the “data transmission interval”, and a timer event that triggers the execution of the failure determination process is issued. Timer TMR is set as follows. Of course, as the margin interval, an arbitrary value can be set within a range in which the reception timing and the determination timing are not different from each other.

  Also, when using a data reception asynchronous timer, for example, the sensor node failure determination process execution interval (that is, timer event issue interval) is set to the “data transmission interval” or “data” set in the sensor node. A rule is set in advance so that it becomes an arbitrary numerical multiple interval such as 1/2 time of transmission interval or 1/3 time, and a timer event that triggers fault determination processing is issued as such. Set the timer to.

  Further, when the intermittent operation interval required for each sensor node or type differs, the execution timing of the failure determination process of each sensor node may be used properly. In addition, when the observation data request frequency is changed from the application, or when the system administrator requests a change in the intermittent operation interval of the sensor node, the sensor network management server SNMS sets the execution timing of the failure determination process to the operation timing. Change to follow the change in interval. Specifically, the event processing engine EVE receives an operation interval change operation result based on the autonomous timing of the sensor node in response to an operation interval change instruction for a certain sensor node, thereby changing the operation interval at the sensor node. Immediately after confirming the completion of the operation, the timer TMR is set to change the timer event issuance timing.

  Therefore, according to the present invention, the execution timing of the failure determination processing of the sensor node can be optimized, and actual failure detection can be effectively realized. Therefore, it contributes to the realization of efficient system operation and maintenance, and the quality of services provided using the system can be kept high.

(Explanation of FIGS. 4 and 5)
4 and 5 are sequence diagrams when the transmission interval is changed before and after application of the present invention in a system using a data reception asynchronous timer. 4 and 5, the sensor node SNx corresponds to one of the sensor nodes SN1 to SNn in FIG. 1, and GW indicates a base station under the control of the sensor node SNx. 4 and FIG. 5, the configuration management & event processing & command management unit (CFM & EVE & CMM) and the request management SSM are the sensor network configuration management unit CFM, event processing engine EVE, sensor network command management unit in FIG. This corresponds to the session management unit SSM. Further, the management client SNMC in FIGS. 4 and 5 corresponds to the sensor network management client SNMC in FIG.

  FIG. 4 shows that the determination criterion is changed in the sensor network management server SNMS in accordance with a scenario in which the operation interval (“data transmission interval”) of the sensor node SNx is changed from t0 to t1 (where t0> t1). The next data reception time is estimated based on the changed operation interval), but the timer interval for executing the determination process itself (timing for issuing a timer event that triggers the determination process start) has been changed. It shows the operation sequence of the system when there is not. As shown in FIG. 4, issuance of a timer event from the timer TMR occurs asynchronously with reception of measurement data.

  As shown in FIG. 4, for example, when the transmission interval change command for the sensor node SNx is requested from the management client SNMC to the sensor network management server SNMS, the transmission interval change command passes through the sensor network command management unit CMM. Then, it is once transferred to the base station GW having the sensor node SNx. At this time, to which base station GW of the plurality of base stations GW the command should be transferred is solved in the sensor network configuration management unit CFM in the sensor network management server as already described in the description of FIG. The transfer destination base station is transferred from the sensor network command management unit CMM, and the sensor network command management unit CMM holds and manages the transferred commands. The transmission change command transferred to the base station GW is temporarily stored there. This is because the sensor node SNx operates intermittently and cannot receive data at an arbitrary timing. The sensor node SNx can be in a receiving state only immediately after performing intermittent operation and transmitting measurement data. Therefore, the transmission interval change command temporarily stored in the base station GW is transmitted to the sensor node SNx immediately after the measurement data is transmitted during the intermittent operation of the sensor node SNx. The sensor node SNx changes its own operation parameter based on the received transmission interval change command. Specifically, the operation interval is changed based on the transmission interval change command received by the wireless communication control unit RFCTL. Although not shown in FIG. 1, the sensor node always supplies power only to elements having a clock function such as a clock and an RTC (Real Time Clock), and can control the operation interval of intermittent operation with a small amount of power. it can.

  The sensor node SNx that has changed the operation interval returns the interval change result to the base station GW via the wireless communication control unit RFCTL. The event processing unit EVE in the sensor network management server SNMS transfers the notification of the result of the transmission interval change command at the sensor node SNx to the sensor network command management unit via the base station GW. When the sensor network command management unit CMM confirms that the command is completed, the value of the item “data transmission interval” regarding the operation parameter of the sensor node SNx managed by the sensor network configuration management unit CFM is received from t0. Change to t1. As a result, at least the determination criterion for the failure to reach the measurement data (assumed “data transmission interval”) is updated, but the timer interval by the timer TMR in the sensor network management server SNMS is not changed. Even though the transmission interval from the node is shortened, the failure determination is executed only at the conventional interval, that is, at the frequency.

  At this time, as illustrated in FIG. 4, when d4 and d5 of the measurement data transmitted after the sensor node SNx changes the operation interval does not reach the base station GW due to some trouble, Failure determination is not performed until a timer event that triggers the determination occurs. That is, in the case of FIG. 4, the failure detection delay time becomes large, and moreover, failure detection for the measurement data d4 that has not arrived is substantially not completed. Further, in the example of FIG. 4, the failure detection of another measurement data d7 that has not arrived is not completed.

  FIG. 5 shows a case where the judgment criteria and the timer interval are changed by applying the present invention in a scenario in which the operation interval (“data transmission interval”) of the sensor node SNx is changed from t0 to t1 (where t0> t1). The rough operation sequence of the entire system is shown. As shown in FIG. 5, issuance of a timer event from the timer TMR occurs asynchronously with reception of measurement data, as in FIG.

  As a premise, in this embodiment, it is assumed that the event processing unit EVE has previously set a rule that the interval for executing the failure determination processing is the same as the “data transmission interval”. That is, when the measurement data transmission interval from the sensor node is changed due to a change in the measurement frequency request of the measurement data, the timing at which the event processing unit EVE executes the failure determination processing of the sensor node is changed in conjunction. To do.

  As shown in FIG. 5, by applying the present invention, the event processing unit EVE is an operation parameter that is managed by the configuration management unit CFM when an event in which the transmission interval change command for the sensor node SNx is completed is received. The value of “data transmission interval” of the sensor node SNx is changed from t0 to t1. Further, the event processing unit EVE resets the timer to the timer TMR at a time interval determined in accordance with the above-described rules for the timer event issuance interval (c1 in FIG. 5). In the example of FIG. 5, since the condition that the timer event issuance interval is the same as the “data transmission interval” is set, the timer event is issued at the same interval as the newly set transmission interval t1 for the sensor node SNx. The timer is reset. As shown in FIG. 5, by applying the present invention, measurement data d4, d5, and d7 that have increased failure detection time or failed to detect failure are detected delay times. However, it is possible to detect failures reliably.

  For example, by setting the failure detection interval as short as possible in advance, it is possible to shorten the failure detection delay to some extent or the failure detection delay time. However, it is impossible to predict how much transmission interval will be set in the future, and setting a timer event to be issued at a fairly short fixed interval in advance will waste computer resources. Even the whole system becomes inefficient operation. On the other hand, by dynamically adjusting the drive timing of the failure detection determination process of the present invention, even when the transmission interval of measurement data changes dynamically, efficient failure detection that is always synchronized with the interval can be performed. . Even when the data transmission intervals of a plurality of sensor nodes are changed, basically, a common rule can be applied, and only one rule needs to be held or maintained.

  In the above embodiment, the interval is set to be the same as the data transmission interval. However, if the timing for performing the abnormality determination in the present invention is determined by multiplying the predetermined interval by an arbitrary number, there is no waste. The merit of performing abnormality determination at the optimum timing can be obtained. As an arbitrary number, if it is desired to make a more precise determination, it may be a decimal number. If an integer is set, an abnormality determination is always performed once in a plurality of sensing times.

(Explanation of FIGS. 6 and 7)
6 and 7 are sequence diagrams when changing the transmission interval before and after application of the present invention in a system using a data reception synchronous timer. 6 and 7, the constituent elements of the sensor node SNx, timer TMR, configuration management & event processing unit (CFM & EVE), request management SSM, and management client SNMC are associated with FIG. 1 in the same manner as in FIGS. It is done.

  FIG. 6 illustrates a scenario in which the operation interval (“data transmission interval”) of the sensor node SNx is changed from t0 to t1 (where t0 <t1). In FIG. 6, after the base station GW temporarily stores the transmission interval change command, it is the same as the operation in FIGS. 4 and 5 until it is transmitted to the sensor node SNx and the value of the transmission interval is changed from t0 to t1. .

  In the embodiment of FIG. 6, the issuance timing of the timer event from the timer TMR is synchronized with the reception of the measurement data so that the issuance timing is after the initial “data transmission interval” from the data reception time. The operation interval of the timer is set in advance. However, even if the sensor node is changed from t0 to t1 (t0 <t1) so as to extend the measurement interval (or data transmission interval), the setting of the timer TMR is not changed. For this reason, if d15 of the measurement data after the sensor node SNx changes the operation interval does not reach the base station GW due to some trouble, the first timer event will perform useless failure determination, and the next Failure determination is not performed until a timer event occurs. That is, in the case of FIG. 6, a delay time occurs in detecting the failure. Furthermore, the detection delay leads to a notification delay to the management client SNMC, and in some cases, a countermeasure delay.

  FIG. 7 shows the overall system when the operation interval of the sensor node SNx (“data transmission interval”) is changed from t0 to t1 (where t0 <t1), and the timer setting is changed by applying the present invention. A rough operation sequence is shown. In the present embodiment, a rule is set in advance in the event processing unit EVE so that “if the data transmission interval from the sensor node SNx is changed, the timer TMR is set and changed so that a timer event is issued at the same time interval”. Has been. As shown in FIG. 7, similarly to FIG. 6, the timer is set so that the issuance of the timer event from the timer TMR occurs in synchronization with the reception of the measurement data.

  As shown in FIG. 7, by applying the present invention, when the event processing unit EVE receives a transmission interval change command completion notification (event) for the sensor node SNx, the event processing unit EVE transfers the completion notification to the sensor network command management unit CMM. Then, the value of “data transmission interval” of the sensor node SNx, which is an operation parameter managed as configuration management information by the configuration management unit CFM, is changed from t0 to t1. Further, the event processing unit EVE resets the timer event issuing interval to the timer TMR to a time interval determined according to a predetermined rule (c11). In the example of FIG. 7, in accordance with the rule that the timer interval is the same as the data transmission interval, the timer event is sent at the same interval as the transmission interval t1 newly set for the sensor node SNx, starting from the time when the data is received. Timer TMR is reset to issue. As shown in FIG. 7, the occurrence of a failure detection delay can be suppressed by applying the present invention.

  Therefore, by applying the present invention and dynamically adjusting the drive timing of the failure detection determination process, efficient failure detection is possible even when the transmission interval of measurement data changes dynamically. In addition, since the rules of the present invention do not include constants specific to each sensor node and are defined as variable or attribute types such as transmission intervals, individual settings are not required for each sensor node. Since it can be applied to a plurality of sensor nodes at the same time, there is an advantage that the setting effort is very small.

<Description of FIG. 8>
FIG. 8 shows an attempt to change the operation interval (measurement data transmission interval) of the sensor node SNx from t0 to t1 (however, t0 <t1) in the system using the data reception synchronous timer as in FIG. However, while the operation interval change command is temporarily stored in the base station GW, the measurement data (d23) does not reach the base station GW, and the temporarily stored command cannot be delivered to the sensor node. Is shown.

  FIG. 8 shows an example in which the measurement data (d23 in FIG. 8) after receiving the transmission interval change command has not arrived. A failure of the sensor node is detected in the failure determination process in the event processing unit EVE. In response to the failure detection result, the management item “operating state” of the corresponding sensor node is updated to “failure” in the configuration management unit CFM. The failure detection result is also transferred to the sensor network command management unit CMM.

  The sensor network command management unit CMM holds information indicating what command is issued to which sensor node by a request from which. Therefore, when a failure notification is received from the configuration management unit CFM, the sensor network command management unit CMM issues a transmission interval change command via the base station to the sensor node that is in the “operational state” of “failure”. The process of canceling the issued transmission interval change command is transmitted to the base station GW. At this time, the base station GW discards the transmission interval change command temporarily held.

  In addition, the sensor network command management unit CMM notifies the request management unit SSM (command execution direct request source viewed from the sensor network command management unit CM) that the command execution result is a failure. Further, the request management unit SSM notifies the management client SNMC that the command execution has failed. The request management unit SSM manages the request source of the request even when there are a plurality of clients corresponding to the management client SNMC.

  As a slightly different case, although not shown, when the sensor network command management unit CMM receives a transmission interval change command, the configuration manager CFM is inquired and the “operating state” of the target sensor node is already “ If it is in the “failure” state, the command is not transferred to the sensor node via the base station GW, and a response indicating that “the command cannot be executed due to the failure” is returned to the request management unit SSM. . As a result, the request management unit SSM similarly notifies the request management client SNMC that the command cannot be executed when a failure occurs as a response to the command request.

  When the sensor node has a function of autonomously changing the “data transmission interval”, an event for notifying that the data transmission interval has been changed may be transmitted. By processing the change notification event in the same manner as the transmission interval change command completion notification, the execution timing of the failure determination process can be adjusted in response to the autonomous data transmission interval change of the sensor node.

  Here, the result notification of the operation interval change command in the sensor node SNx or the data transmission interval change notification event from the sensor node that autonomously changed the data transmission interval is not transmitted to the sensor network management server SNMS for some reason. Cases can happen. Although the probability of occurrence is low, in this case, the sensor network management server SNMS cannot confirm any change in the operation interval of the sensor node. In such a case, for example, in the sensor network management server SNMS, the configuration management unit CFM holds two or more consecutive data and their data reception times transmitted from the corresponding sensor node SNx as much as possible, and By having the function of calculating the data reception interval from the information, the operation interval after the change of the sensor node SNx can be estimated. By using the result for resetting the execution timing of the failure determination process, the failure determination process can be maintained so as to be executed efficiently.

  Next, an application for dynamically changing the measurement frequency of measurement data acquired by a specific sensor node, that is, the transmission frequency will be described with three specific examples.

  The first example is an application example of air conditioning control. Air conditioning control can be performed by monitoring indoor temperature and humidity with a sensor node equipped with a temperature and humidity sensor. For example, a human sensor is installed in the sensor node to monitor whether a person is in the room at the same time as the temperature and humidity. During times when there is no person in the room (for example, at night), the frequency of temperature and humidity measurement and data transmission It is conceivable to reduce the temperature (for example, once every 5 minutes) and increase the temperature / humidity measurement and data transmission frequency (for example, once a minute) during the time zone (daytime) when a person is present in the room. In the daytime, not only does the person receive the air conditioning service, but also increases the probability that the person will become a shield for wireless communication. Therefore, in order to maintain the quality of wireless communication, the daytime is more than the frequency required by the application. There is also an effect of ensuring the reliability of wireless communication by increasing the transmission frequency of temperature and humidity data, that is, shortening the operation interval of the sensor node. Specifically, it is determined whether or not the human sensor information received by the event processing engine EVE of the sensor network management server exceeds a predetermined value, and based on the determination result, each sensor node is warmed as necessary. Outputs a command to change the humidity data transmission interval.

  The second example is a fire monitoring system. When there are no people in the building, relax the temperature monitoring (long measurement cycle). Battery life can also be extended by reducing the frequency of temperature monitoring. For example, the temperature is monitored at intervals of 5 minutes, and the allowable delay time for failure detection is set to a maximum of 5 minutes (that is, the failure determination processing execution interval). On the other hand, when there are people in the building, temperature monitoring and system failure monitoring are strictly performed because it is also related to human life. For example, the temperature is monitored at intervals of 30 seconds, and the allowable delay time for failure detection is a maximum of 10 seconds (as fast as possible). As in the first example, a human sensor is used to determine whether or not a person is in the building, and the sensor data management is used to control the temperature data transmission interval based on information from the human sensor. It is a server.

  A third example is a health monitoring system. When the pulse, body temperature, etc. are normal, the monitoring interval of various biological information such as the pulse and the maximum allowable delay time for fault monitoring are set loosely to extend the battery life. On the other hand, when an abnormality is detected in a pulse or the like, the monitoring frequency of biological information such as a pulse is increased in order to increase the observation accuracy, and accordingly, the maximum allowable delay time for failure monitoring of the sensor node is also set short. In this example as well, the monitoring frequency (measurement interval) change control is realized as a function of the sensor network management server. In this example, biological information such as a pulse to be measured is added without adding a new sensor. The change timing of the measurement interval, that is, the data transmission interval can be determined based on whether or not an abnormal value is indicated.

  As in the above three examples, you may want to dynamically change the transmission frequency of measurement data according to the time zone and situation. In such a case, the accuracy of fault monitoring (allowable delay time for fault detection) is also included in the transmission interval. It is necessary to loosen and tighten them together.

  As described above, in the sensor network application system, it is assumed that the value of the transmission frequency of measurement data is changed according to various requirements. Based on the tendency of presence or absence of signals (including measurement data such as temperature) periodically transmitted from the sensor node, the sensor node is regarded as a heartbeat signal, and sensor node life / death determination (communication availability determination) is performed. Therefore, it can be said that monitoring the heartbeat signal in consideration of the change of the interval is a new viewpoint unique to the sensor network system and a new problem in monitoring the sensor network system.

(Remote operation monitoring service)
In addition, when a third party performs system operation, monitoring, and maintenance services as described above, it is possible to provide high-quality services to customers while minimizing maintenance work such as battery replacement. Become. For example, in FIG. 1, when the wired network WDN is the Internet or a network using some dedicated line, the user terminal UST is a computer placed in a monitoring service center geographically far from the sensor network system, for example. Can be considered. Then, by operating the sensor network management client SNMC in the user terminal UST at a remote place, it is possible to easily implement a change of the measurement data transmission interval for each sensor node included in the present invention as a remote service.

  As described above, the present invention can be applied to a sensor network system including various types of sensor nodes and its operation and maintenance management service.

1 is a block diagram of a sensor network system showing a first embodiment of the present invention. FIG. 5 is a failure monitoring sequence diagram for measurement data non-delivery when a data reception asynchronous timer is used. FIG. 5 is a failure monitoring sequence diagram for non-delivery of measurement data when a data reception synchronous timer is used. FIG. 7 is a sequence diagram when a failure occurs in failure monitoring without using the present invention after using a data reception asynchronous timer in a scenario involving a transmission interval change request. FIG. 6 is a sequence diagram in a case where a transmission interval change request is used and a failure of fault monitoring is suppressed by applying the present invention after using a data reception asynchronous timer. FIG. 10 is a sequence diagram in another case where failure monitoring occurs when a data reception synchronous timer is used and the present invention is not applied in a scenario involving a transmission interval change request. FIG. 11 is a sequence diagram in another case where failure monitoring is appropriately performed by applying the present invention using a data reception synchronous timer in a scenario involving a transmission interval change request. The sequence diagram when a transmission interval change request is not transmitted to a sensor node. An example of a table indicating configuration management data.

Explanation of symbols

SN1 to SNn Sensor nodes GW1 to GWm Base station SNMS Sensor network management server WDN Wired network WLN Wireless network RFC Wireless circuit PWR Power supply CPU Microcomputer SDEV Sensor device ADEV Actuator device RFCTL Wireless communication control unit PWCTL on sensor node Power supply control on sensor node Unit MCTL operation control unit SCTL on sensor node sensor control unit ACTL on sensor node actuator control unit UPC on sensor node uplink packet conversion unit DPC on base station downlink packet conversion unit GIDM on base station ID management on base station Unit GCMM Command management unit on base station GRFC Wireless communication control unit on base station GNIC Wired communication control unit on base station SSM Session management unit on sensor network management server EVE Event processing engine SCM on the network management server Session management unit EXTIF on the sensor network management server External device cooperation adapter TMR on the sensor network management server Timer CMM on the sensor network management server Sensor network command management unit CFM on the sensor network management server Sensor network configuration management unit DBMS on the sensor network management server Database management unit SNIC on the sensor network management server (wired) communication control unit on the sensor network management server.

Claims (9)

A communication unit that receives data transmitted at predetermined intervals from a terminal connected via a network;
A determination unit that determines presence / absence of the data to be received and performs abnormality determination of the terminal;
The determination unit determines an interval for performing the abnormality determination for the terminal based on the predetermined interval according to a predetermined rule.   2. The data processing system according to claim 1, wherein the rule determines an interval for performing the abnormality determination by multiplying the predetermined interval by an arbitrary number. When the determination unit receives a notification that the interval at which data is transmitted from the terminal is changed,
3. The data processing system according to claim 1, wherein an interval for performing the abnormality determination is re-determined according to the predetermined rule.   4. The data processing according to claim 1, wherein when the data transmitted from the terminal is received, the determination unit counts the passage of the interval for performing the abnormality determination again from the reception time. 5. system. The terminal is a wireless sensor node,
5. The data processing system according to claim 1, wherein the data received by the receiving unit is sensing data that is wirelessly transmitted by the sensor node at a predetermined interval.   6. The determination unit according to claim 1, wherein when the wireless terminal is determined to be abnormal as a result of the determination, the determination unit notifies the connected higher-level device that the wireless terminal is abnormal. The data processing system described in Crab.   The data processing system according to claim 1, wherein an instruction to change the data transmission interval of the terminal is issued by the determination unit and transmitted to the terminal.   8. The data processing system according to claim 7, wherein the determination unit issues an instruction to cancel the change instruction when an operation state of the terminal that is the target of the change instruction is “abnormal”. The determination unit determines an operation interval of the sensor node from a timing at which data is continuously received from the terminal at least twice or more, and adapts the predetermined rule to the determined operation interval to detect an abnormality of the terminal. 9. The data processing system according to claim 1, wherein a timing for performing the determination is determined.